水下图像的多尺度分数列维稳定运动模型  

Multi-Scale Fractional Lévy Stable-Motion Model of Underwater Image

作　　者：李旭涛[1] 金连文[1] 高学[1] 尹明[2] 

机构地区：[1]华南理工大学电子与信息学院,广东广州510640 [2]广东工业大学自动化学院,广东广州510006

出　　处：《华南理工大学学报（自然科学版）》2008年第9期81-85,共5页Journal of South China University of Technology(Natural Science Edition)

基　　金：国家“863”计划项目(2002AA401001-4B);中国博士后科学基金资助项目(20070410823)

摘　　要：分析了水下散射图像的增量分布及自相似特性.针对传统分数布朗运动(FBM)模型存在的不足,基于分数列维稳定运动(FLSM)模型,将纹理表面的自相似指数看作依度量尺度而变化的变量,提出了一种新的多尺度分数列维稳定运动(MFLSM)模型,并将该模型扩展至二维用于对水下散射图像的仿真.仿真结果表明,该模型具有对纹理多尺度自相似性以及散斑纹理的表征能力.This paper analyzes the increment distribution and self-similarity of typical underwater images. In order to overcome the disadvantages of the traditional fractal Brownian motion （FBM） model, a muhi-seale fractional Lévy stable-motion （MFLSM） model is proposed based on the fractional Lévy stable-motion （FLSM） model, in which the self-similarity parameter of image surface is considered as a variable to vary with the measure scale. The model is then extended to a 2D space and is further employed to simulate underwater scattering images. The results show that the proposed MFLSM model represents the multi-scale self-similarity and the speckle of underwater optical images well.

关 键 词：图像处理 散射图像 多尺度自相似性 稳定过程 

分 类 号：TN911.73[电子电信—通信与信息系统]